Relating daily HRV fluctuations to stress & somatisation

Oura ring HRV fluctuations

On January 12th, 2022, Healthcare published our latest peer-reviewed research on Heart Rate Variability (HRV). The paper is titled “Trends in Daily Heart Rate Variability Fluctuations Are Associated with Longitudinal Changes in Stress and Somatisation in Police Officers” and is part of a special issue on Mental and Behavioral Healthcare. In this blogpost, I will attempt to summarize the article and how it complements our prior research in more lay language.


Psychological stress has an evolutionary function to prepare us for a response to cope with the demands that triggered it. However, it can also lead to mental and physical health issues and absenteeism if sustained over a longer period with limited recovery opportunities. Due to the emergence of wearable sensors, monitoring physiological signs of stress and its impact is getting increasingly feasible. My PhD project, including this study, focuses on exploring if wearables can be used to monitor physiological or behavioral signs of changes in the mental resilience of employees with a stressful job. If successful, those insights may be used in interventions that provide an early warning signals if there are signs of stress accumulation, which could help to limit the negative impact of stress.

Heart Rate Variability (HRV)

After review of scientific literature, resting HRV has become a main focus of this research. HRV is a measure for the amount in which our heart rate tends to accelerate and slow down between beats. In general, a higher resting HRV is often linked to more favorable health outcomes and performance, whereas a lower HRV may be related to mental or physical stress.

For example, in our first study, we found that interns that woke up with a higher-than-normal resting HRV reported less stress on a demanding day and less mental exhaustion on a stressful day. They also woke up with a slightly lower resting HRV after feeling mentally exhausted the prior evening. Therefore, waking up with a high resting HRV appeared to show that these interns were more resilient that day, whereas mental exhaustion tends to drain those resources.

Most of the existing research on this topic looks at differences in the HRV between individuals. This research helps us understand why the HRV is higher in some individuals than in others, but needs to be complemented by research that assesses differences within individuals to help us understand what causes changes in our own HRV over time. Our first study did this by assessing within-day differences in HRV, whereas our latest study attempts to expand on this by exploring longitudinal (5-week) trends in the resting HRV.

Fluctuations in the daily HRV

Besides trends in resting HRV, our latest study also explored if the amount of fluctuations in the daily HRV may be related to mental well-being. Literature from the sports science field suggests that an increase in the amount of day-to-day variation in the HRV can be related to increased fatigue, stress or a decreased stress tolerance in athletes. Put differently, when the day-to-day changes in the HRV become increasingly extreme (either abnormally high and/or low), it may be related to an increase in stress-related outcomes. This body of knowledge is novel in sports sciences, and to our knowledge still unexplored in the field of psychology prior to our study.

What we did

Between June 2020 and July 2021, 10 police officers were recruited to participate, of which 9 were included in this analysis (1 was excluded due to diagnosed heart arrhythmias). The police officers wore an Oura ring, a consumer-available wearable that measures HRV during sleep, as well as sleep (quality), body temperature, heart rate and daytime physical activity. Additionally, they filled in a short daily evening questionnaire and a longer one every 5 weeks that measured stress, depression, anxiety and somatisation.

In our analyses, we estimated the 5-week HRV trend by doing a linear regression analysis over the daily HRV measures. A linear regression analysis can be visualized as the straight line through all collected data points that best represents the trend in the data. The figure below is an example of this (which is unrelated to our study, but a helpful visualization to explain what linear regression is). The steeper the red line increases (the ‘slope’), the more positive the trend through the data is. If the red line decreases, there is a negative trend in the data.

An example of a linear regression analysis (source)

The daily HRV fluctuations were quantified by calculating the standard deviation for each datapoint based on the previous week. This is called a ‘7-day rolling standard deviation’, and referred to as HRVsd in our paper and the figures down below. Finally, changes in stress, depression, anxiety and somatisation were quantified by simply calculating the difference between the start and the end of each 5-week period.

The 9 included participants collected HRV data on 1648 unique person-days and filled in 57 longitudinal questionnaires. Taken together, we were able to compare trends in HRV with questionnaire data on 47 periods, which formed the sample size of our study.

Daily HRV fluctuations are related to changes in stress

As expected based on the sports science literature, a positive trend in the 5-week in the amount of day-to-day fluctuations in HRV could be related to an increase in stress. Trends in the daily HRV itself were not associated with changes in stress. Our model, which also included trends in the total sleep time, physical activity and alcohol consumption but were not significant predictors, explained 18.5% of the variance in changes in stress. This means that changes in daily HRV fluctuations are a relevant predictor of changes in stress, but that there are still other factors at play as well.

The figure down below shows that when the HRV fluctuations (HRVsd) trended up (right half of the x-axis), stress tended to increase (upper half of the y-axis). On the other hand, when the amount of fluctuations decreased (left half of the y-axis), stress tended to decrease (bottom half of the y-axis).

Daily HRV fluctuations are also related to somatisation

An increasing amount of fluctuations in the daily HRV was also linked to increased somatisation, which are bodily symptoms that may be related to stress. However, we observed something else here as well. Increasing fluctuations in the daily HRV were only associated with increasing somatisation if the daily HRV itself trended down or was neutral, but if the daily HRV itself increased during the same period. This means that in principle, the increasing daily HRV fluctuations were often related to increased somatisation – unless the underlying system itself ‘improved’ regardless. This specific situation can be seen as an example of a resilience or even anti-fragility. The full model explained 21.3% of the variance in changes in stress.

The figure down below (an interaction plot) is a bit more complex than the previous ones, as the y-axis does not represent a single variable, but the relationship between daily HRV fluctuations (HRVsd) and somatisation. The x-axis represents the trend in the underlying HRV. The gray area around the declining black line shows the confidence interval of the regression. When this confidence interval is above zero (the horizontal striped line), there is a statistically significant effect. When the zero-line crosses this confidence interval, there is no significant effect. As can be seen, when the HRV uptrend is larger than circa 0.2, the relationship between HRV fluctuations and somatisation is no longer significant.

Floor effects for depression and anxiety

We also did these analyses for depression and anxiety, but found no relationships there. However, there was a good explanation for that. In total, 92.9% of the questionnaires scored ‘0’ on depression and 87.5% scored ‘0’ on anxiety. None of these questionnaires scored above the thresholds for ‘moderately’ or ‘severely’ elevated levels, illustrating that there was a complete absence of clinically relevant symptoms. Without variation in the data, you do not find statistically significant relationships. Therefore, it is possible that in a population with more symptoms of depression and anxiety, such a relationship still might exist. Future studies could consider using more sensitive questionnaires for depression and anxiety.

What we learned

When the physiological balance of these police officers was increasingly challenged (increasing daily HRV fluctuations) over time, they tended to report more stress and somatic symptoms. However, if their underlying physiological system itself actually appeared to get stronger despite those challenges (increasing daily HRV itself), they did not necessarily develop somatic symptoms. Based on these findings, monitoring trends in daily HRV and fluctuations therein may be useful to help detect changes in stress and somatisation in police officers over time.

Since this was (to our knowledge) the first study to apply these novel methods in this population and field based on a modest sample, confirmation in future studies is needed to increase confidence in the presented findings. The fact that these findings are in line with those of prior research in the field of sport science can be seen as a sign that this is a topic that deserves further investigation.

Leave a comment

Your email address will not be published. Required fields are marked *